ergentOrder.onnx-scala-backends_2.13.0.11.0.source-code.ORTOperatorBackendAll213.scala Maven / Gradle / Ivy
The newest version!
package org.emergentorder.onnx.backends
import org.emergentorder.onnx._
//Going forward we only support ops (and their versions) which are supported in both ONNX Runtime & ONNX.js (CPU), opset 6+
//Plus a handful of others we need
//Current ONNX opset 12, no new ops in opset 13, just support for training
class ORTOperatorBackendAll
extends ORTOperatorBackend
with AbsV6
with AcosV7
with AcoshV9
// with AdagradV1 //New in 1.7.0, training
with AddV7
with AndV7
with ArgMaxV11
// with ArgMinV11 //Strangely missing from ONNX.js
// with ArrayFeatureExtractorV1 //ONNX ML, not tested for in scoreboard
with AsinV7
with AsinhV9
with AtanV7
with AtanhV9
with AveragePoolV7
with AveragePoolV10
// with AveragePoolV11 //Missing from ONNX.js
with BatchNormalizationV7
with BatchNormalizationV9
// with BinarizerV1 //ONNX ML, not tested for in scoreboard
// with BitShiftV11 //not supported in ONNX.js
with CastV9
// with CastMapV1 //ONNX ML, not tested for in scoreboard
// with CategoryMapperV1 //ONNX ML, not tested for in scoreboard
// with CeilV1
with CeilV6
// with CeluV12 //new in 1.7.0
with ClipV6
// with ClipV11 //not supported in ONNX.js
// with CompressV11
with ConcatV4 //Retained for Squeezenet v1.1
with ConcatV11
// with ConcatFromSequenceV11
with ConstantV11
// with ConstantOfShapeV9
with ConvV1 //Retained for Squeezenet v1.1
with ConvV11
// with ConvIntegerV10
// with ConvTransposeV11
with CosV7
with CoshV9
// with CumSumV11
// with DepthToSpaceV1
// with DequantizeLinearV10
// with DetV11
// with DictVectorizerV1 //ONNX ML, not tested for in scoreboard
with DivV7
with DropoutV7
with DropoutV10
// with DynamicQuantizeLinearV11
// with EinsumV12 //new in 1.7.0
with EluV6
with EqualV11 //Missing from ONNX.js, but we need it
// with ErfV9
with ExpV6
with ExpandV8
// with EyeLikeV9
// with FeatureVectorizerV1 //ONNX ML, not tested for in scoreboard
with FlattenV11
with FloorV6
// with GRUV7
with GatherV11
// with GatherElementsV11
// with GatherNDV11
with GemmV11
with GlobalAveragePoolV1
// with GlobalLpPoolV2 //fails in scoreboard
with GlobalMaxPoolV1
// with GradientV1 //Training, new in 1.7.0
// with GraphCallV1 //Training, new in 1.7.0
with GreaterV9 //Missing in ONNX.js, but we need it
with GreaterOrEqualV12 //Missing in ONNX.js, but we need it
// with HardSigmoidV6
// with HardmaxV11
// with IdentityV1
// with IfV11 //fails in scoreboard
// with ImputerV1 //ONNX ML, not tested for in scoreboard
with InstanceNormalizationV6
// with InverseV12 //New in 1.7.0
// with IsInfV10
with IsNaNV9
with LRNV1
// with LSTMV7
// with LabelEncoderV2 //ONNX ML, not tested for in scoreboard
with LeakyReluV6
with LessV9 //Missing in ONNX.js, but we need it
with LessOrEqualV12 //Missing in ONNX.js, but we need it
// with LinearClassifierV1 //ONNX ML, not tested for in scoreboard
// with LinearRegressorV1 //ONNX ML, not tested for in scoreboard
with LogV6
// with LogSoftmaxV11
// with LoopV11
// with LpNormalizationV1 //fails in scoreboard
// with LpPoolV11 //fails in scoreboard
with MatMulV9
// with MatMulIntegerV10
with MaxV8 //Fails in ONNX.js, but we need it
with MaxPoolV1 //Retained for Squeezenet v1.1
// with MaxPoolV11 //Missing in ONNX.js
// with MaxRoiPoolV1 //fails in scoreboard
// with MaxUnpoolV11
// with MeanV8 //Missing in ONNX.js, we can use ReduceMean instead
// with MeanSquaredDistanceV12
// with MeanVarianceNormalizationV9
with MinV8 //Missing in ONNX.js, but we need it
with ModV10 //Missing in ONNX.js, but we need it
// with MomentumV1 //Training, new in 1.7.0
with MulV7
// with MultinomialV7 //fails in scoreboard
with NegV6
// with NegativeLogLikelihoodLossV12 //new in 1.7.0
// with NonMaxSuppressionV11
// with NonZeroV9
// with NormalizerV1 //ONNX ML, not tested in scoreboard
with NotV1
// with OneHotV11
// with OneHotEncoderV1 //ONNX ML, not tested in scoreboard
with OrV7
with PReluV9
with PadV11
with PowV7
// with QLinearConvV10
// with QLinearMatMulV10
// with QuantizeLinearV10
// with RNNV7
// with RandomNormalV1 //fails in scoreboard
// with RandomNormalLikeV1 //fails in scoreboard
// with RandomUniformV1 //fails in scoreboard
// with RandomUniformLikeV1 //fails in scoreboard
with RangeV11 //Missing in ONNX.js, but we need it
with ReciprocalV6
// with ReduceL1V11
// with ReduceL2V11
with ReduceLogSumV11
// with ReduceLogSumExpV11
with ReduceMaxV11
// with ReduceMeanV1
with ReduceMeanV11
with ReduceMinV11
with ReduceProdV11
with ReduceSumV11
with ReduceSumSquareV11
with ReluV6
with ReshapeV5
// with ResizeV11
// with ReverseSequenceV10
// with RoiAlignV10
with RoundV11 //Missing in ONNX.js, but we need it
// with SVMClassifierV1 //ONNXML, not tested for in scoreboard
// with SVMRegressorV1 //ONNXML, not tested for in scoreboard
// with ScalerV1 //ONNXML, not tested for in scoreboard
// with ScanV11
// with ScatterV11 //Deprecated in 1.8
// with ScatterElementsV11
// with ScatterNDV11
// with SeluV6
// with SequenceAtV11
// with SequenceConstructV11
// with SequenceEmptyV11
// with SequenceEraseV11
// with SequenceInsertV11
// with SequenceLengthV11
with ShapeV1
// with ShrinkV9
// with SigmoidV1
with SigmoidV6
with SignV9
with SinV7
with SinhV9
// with SizeV1
with SliceV11
// with SoftmaxV1
with SoftmaxV11
// with SoftmaxCrossEntropyLossV12 //new in 1.7.0
// with SoftplusV1
// with SoftsignV1
// with SpaceToDepthV1 //fails in scoreboard
// with SplitV2
// with SplitV11 //Nice to have
// with SplitToSequenceV11
with SqrtV6
with SqueezeV11
// with StringNormalizerV10
with SubV7
with SumV8
with TanV7
with TanhV6
// with TfIdfVectorizerV9
// with ThresholdedReluV10
with TileV6
// with TopKV11 //Nice to have
with TransposeV1
// with TreeEnsembleClassifierV1 //ONNX ML, not tested for in scoreboard
// with TreeEnsembleRegressorV1 //ONNX ML, not tested for in scoreboard
// with UniqueV11
// with UnsqueezeV1
with UnsqueezeV11
// with UpsampleV10 //Deprecated in 1.8
// with WhereV9
with XorV7
// with ZipMapV1 //ONNX ML, not tested for in scoreboard
© 2015 - 2025 Weber Informatics LLC | Privacy Policy